Open Access
Depression among patients with
To cite: Ambaw F,
Mayston R, Hanlon C, et al.
Depression among patients
with tuberculosis:
determinants, course and
impact on pathways to care
and treatment outcomes in a
primary care setting in
southern Ethiopia—a study
protocol. BMJ Open 2015;5:
e007653. doi:10.1136/
bmjopen-2015-007653
▸ Prepublication history for
this paper is available online.
To view these files please
visit the journal online
(http://dx.doi.org/10.1136/
bmjopen-2015-007653).
Received 12 January 2015
Revised 20 March 2015
Accepted 10 April 2015
For numbered affiliations see
end of article.
Correspondence to
Fentie Ambaw;
fentambaw@yahoo.com
tuberculosis: determinants, course and
impact on pathways to care and
treatment outcomes in a primary care
setting in southern Ethiopia—a study
protocol
Fentie Ambaw,1,2 Rosie Mayston,3 Charlotte Hanlon,2,3 Atalay Alem2
ABSTRACT
Introduction: Depression is commonly comorbid with
chronic physical illnesses and is associated with a
range of adverse clinical outcomes. Currently, the
literature on the role of depression in determining the
course and outcome of tuberculosis (TB) is very
limited.
Aim: Our aim is to examine the relationship between
depression and TB among people newly diagnosed and
accessing care for TB in a rural Ethiopian setting. Our
objectives are to investigate: the prevalence and
determinants of probable depression, the role of
depression in influencing pathways to treatment of TB,
the incidence of depression during treatment, the
impact of anti-TB treatment on the prognosis of
depression and the impact of depression on the
outcomes of TB treatment.
Methods and analysis: We will use a prospective
cohort design. 703 newly diagnosed cases of TB (469
without depression and 234 with depression) will be
consecutively recruited from primary care health
centres. Data collection will take place at baseline, 2
and 6 months after treatment initiation. The primary
exposure variable is probable depression measured
using the Patient Health Questionnaire-9. Outcome
variables include: pathways to treatment, classical
outcomes for anti-TB treatment quality of life and
disability. Descriptive statistics, logistic regression and
multilevel mixed-effect analysis will be used to test the
study hypotheses.
Ethics and dissemination: Ethical approval has
been obtained from the Institutional Review Board
(IRB) of the College of Health Sciences, Addis
Ababa University. Findings will be disseminated
through scientific publications, conference
presentations, community meetings and policy
briefs.
Anticipated impact: Findings will contribute to a
sparse evidence base on comorbidity of depression
and TB. We hope the dissemination of findings will
raise awareness of comorbidity among clinicians and
service providers, and contribute to ongoing debates
Ambaw F, et al. BMJ Open 2015;5:e007653. doi:10.1136/bmjopen-2015-007653
Protocol
BMJ Open: first published as 10.1136/bmjopen-2015-007653 on 8 July 2015. Downloaded from http://bmjopen.bmj.com/ on May 4, 2022 by guest. Protected by copyright.
regarding the delivery of mental healthcare in
primary care in Ethiopia.
INTRODUCTION
The relationship between depression and
chronic physical illnesses is bidirectional.
Comorbidity is associated with a range of
adverse outcomes, including functional
impairment, increased medical costs, poor
adherence to medication and self-care regi-
mens, increased medical symptom burden
and increased mortality.1 2 Individually,
depression3 4 and tuberculosis (TB)4 5 are
recognised as important public health con-
cerns, contributing to 2.5–2.0% of disability
adjusted life years (DALYs) worldwide in
2010, respectively.6 In Ethiopia, the preva-
lence of depression was found to be 9.1% in
a nationally representative sample;7 and in a
population-based survey carried out in south-
ern Ethiopia, depression was found to be the
seventh leading cause of disease burden con-
tributing to 6.5% of the DALYs in 1998.8 TB
is a key public health concern in Ethiopia: in
2009/2010, it was the second most important
cause of death.9 Ethiopia is ranked seventh
among the 22 high-burden countries that
account for 81% of all cases of TB and 80%
of all TB deaths worldwide. Ethiopia is also
one of 27 countries identified as having a
high prevalence of multidrug resistant TB
(MDR-TB). The burden of MDR-TB in these
countries accounts for 86% of cases
worldwide.9 10
Evidence from cross-sectional studies
(some of which were carried out in hospital
1
Open Access
settings in African countries) indicate a very high preva-
lence of comorbid depression (ranging from 10% to
52%) among patients with TB.11–18 However, because
longitudinal research on TB and depression is scarce,
the nature of the relationship and trajectory of
comorbidity is little understood. Most high-quality
studies examining the prevalence and impact of
comorbid depression in the context of chronic physical
diseases were conducted among people with diabetes
mellitus, ischaemic heart diseases, cancer and chronic
obstructive pulmonary diseases.1 2 19 The extent to
which the impact of comorbid depression in the context
of TB is comparable to depression comorbid with other
non-communicable chronic conditions is unclear. TB is
a curable condition with relatively shorter treatment dur-
ation and therefore has the potential for more favour-
able outcomes than many non-communicable chronic
diseases.16 However, TB remains a debilitating, stigma-
tised communicable disease requiring complex and
aggressive treatment. Similar to depression in the
context of HIV/AIDS, the potential for depression to
impair adherence to complex TB medication regimens
is not only problematic in terms of individual patient
outcomes but also poses a threat to public health
through the potential for the development of multidrug
resistance.20 For now, the literature on the role of
depression in determining TB treatment outcomes
(treatment default, interruption, completion, failure,
death) and treatment pathways (routes of help seeking
for TB treatment within modern or traditional care
systems) is very limited.
POSSIBLE MECHANISMS FOR THE ASSOCIATION
BETWEEN TB AND DEPRESSION
It is likely that pathways for associations between TB and
depression are complex and multidirectional. Biological
and psychosocial pathways may be responsible for observed
associations. The extent to which different pathways con-
tribute to the burden of comorbidity is currently unclear.
For example, some researchers have suggested that patients
with TB may develop depression as a result of chronic
infection or related psychosocioeconomic stressors21 or
due to the effects of treatment such as isoniazid.22 An alter-
native pathway may be that TB is contracted as a result of
compromised immunity and neglected self-care associated
with depression.23 Finally, there is evidence to suggest that
TB and depression may share risk factors.2 23 24
Immunological responses have been implicated in the
association between chronic disease and depression.
Chronic infectious conditions may lead to overproduc-
tion of proinflammatory cytokines such as interleukin 6,
which facilitate cascades of endocrine reactions that are
suggested to result in depressive symptoms.24 However,
there is growing evidence that depression itself enhances
the production of proinflammatory cytokines and dir-
ectly minimises the immunological competence of
patients by downregulating cellular and humoral
2 Ambaw F, et al. BMJ Open 2015;5:e007653. doi:10.1136/bmjopen-2015-007653
responses.1 2 23 24 In addition, in chronic pulmonary
BMJ Open: first published as 10.1136/bmjopen-2015-007653 on 8 July 2015. Downloaded from http://bmjopen.bmj.com/ on May 4, 2022 by guest. Protected by copyright.
conditions with hypoxia, the hypoxic condition can act
directly to make patients anxious and depressed.
Likewise, general factors associated with chronic disease
such as weight loss, fatigue, psychological and social
losses may trigger depressive reactions.21
There is a consensus among experts that people who
have chronic diseases and comorbid depression can
benefit from treatments for depression, including treat-
ment with antidepressants.1 14 19 25 26 If depression in the
context of TB is associated with a biological pathway or is
a response to the burden of chronic infection, it might
be expected that treatment of TB may lead to reduced
symptoms of depression, perhaps without the need for
further intervention. As intervention for depression
among patients with TB is likely to incur additional costs,
pill burden and potential stress, it is important to under-
stand to what extent TB treatment alone may be an effect-
ive intervention for depressive symptoms. In addition,
anti-TB medications, especially isoniazid, the first of the
monoamine oxidase inhibitors to be considered for the
treatment of mental disorders in the 1950s,27 and now a
core drug in anti-TB treatment,28 may have significant
interactions with selective serotonin reuptake inhibi-
tors,27 which are WHO-recommended drugs for the treat-
ment of depression in the mental health gap action
programme (mhGAP) guideline.29 The integration of
mental healthcare into primary care settings is currently
being scaled up in Ethiopia,29 30 with depression as one
of the priority disorders. The results of the proposed
study will help to inform the targeting and delivery of
mental health services in the context of TB in Ethiopia.
AIMS AND OBJECTIVES
Our overall aim is to carry out a longitudinal study of
depression in the context of TB, in order to determine
the impact of comorbid depression on TB outcomes. In
this way, we hope to contribute to a sparse evidence base
that lacks high-quality evidence from African countries,
where the highest rates of cases and deaths relative to
population size occur.31 Our study has five objectives:
1. To determine the prevalence of depression in people
with TB at the time of anti-TB treatment initiation,
2. To assess factors associated with baseline depression,
3. To determine the incidence (risk) of depression in
patients with TB at 2 and 6 months after starting
anti-TB treatment,
4. To assess the impact of depression on anti-TB treat-
ment outcomes (classical treatment outcomes: treat-
ment default, interruption, completion, failure,
death, disability score and health-related quality of
life) at 2 and 6 months, and to explore the moderat-
ing effect of mhGAP depression interventions deliv-
ered in routine settings,
5. To assess whether depression is associated independ-
ently with longer pathways to anti-TB treatment
(after adjusting for sociodemographic variables).

HYPOTHESES
1. People with TB who have depression at the
time-of-treatment initiation (baseline) will have worse
treatment outcomes of TB (classical treatment out-
comes, disability score and health-related quality of
life) at the end of 2 and 6 months follow-up when
compared with those without depression at baseline.
2. Anti-TB treatment will progressively reduce depres-
sive symptoms so that those with depression at base-
line will have reduced severity of depression (Patient
Health Questionnaire-9; PHQ-9 scores) or no depres-
sion after 2 and 6 months treatment for TB.
METHODS AND ANALYSIS
Study setting
The study will be conducted from December 2014 to
December 2015 in nine primary care facilities in Butajira
town, Mareko district, Meskan district, Sodo district and
Silte district of the Southern Nations, Nationalities and
Peoples’ Region (SNNPR) of Ethiopia (figure 2). Farming
is the main economic activity in the area. In 2012/2013,
there were 2742 people with TB in the zone. Directly
observed treatment short-course (DOTS) for TB is being
implemented in all health facilities.9 In 2011, the anti-TB
treatment defaulter rate, death rate and treatment success
rate were 2.5%, 2.0% and 82.3%, respectively, in SNNPR.32
Delivery of anti-TB treatment and care in rural Ethiopia
The DOT for new patients with TB lasts for 6 months,
and consists of two phases: intensive and continuation.
The intensive phase consists of treatment with a combin-
ation of four medications (rifampicin, ethambutol, iso-
niazid and pyrazinamide) for the first 2 months, and the
continuation phase consists of a combination of two med-
ications (rifampicin and isoniazid), to be taken for
4 months immediately after the intensive phase. A health
worker or community-based anti-TB treatment supporter
has to observe the patient swallow the medications once a
day. Currently, anti-TB treatment is being delivered in
health centres, hospitals (by health workers) and at
health posts (by trained community-based ‘health exten-
sion workers’).9 Health posts are the lowest level of
healthcare in Ethiopia, serving 5000 people. The flow of
patients with TB in each of the health facilities selected
for this study is above 5/month (figures 1 and 2).
Study design
The study is a prospective cohort in which adults with
newly diagnosed TB will be recruited at the time of initi-
ating treatment and followed up to the end of treatment
(6 months after initiation) (figure 1). Data collection
will occur at three time-points: baseline (before anti-TB
treatment), at 2 months (end of intensive phase) and at
6 months (end of continuation phase; table 1). Figure 3
shows the approach to investigation of the effect of TB
on depression and figure 4 shows the approach to inves-
tigation of the effect of depression on TB.
Ambaw F, et al. BMJ Open 2015;5:e007653. doi:10.1136/bmjopen-2015-007653
Open Access
Inclusion criteria:
▸ Patients attending the selected health centres who
BMJ Open: first published as 10.1136/bmjopen-2015-007653 on 8 July 2015. Downloaded from http://bmjopen.bmj.com/ on May 4, 2022 by guest. Protected by copyright.
have started their anti-TB treatment within the last
1 month
▸ Those aged over 17 years.
Exclusion criteria:
▸ Patients with a known plan to be transferred out of
the study sites
▸ Those too ill to be interviewed at baseline as per-
ceived by the interviewer or the patient
▸ Those patients admitted to the in-patient unit for
more than 5 days in the last 1 month
▸ Patients with MDR-TB, who constitute a different popu-
lation because their treatment is different (more toxic
medications for a much longer duration). MDR-TB is a
much more feared and stigmatised condition with
patients probably told that no further treatment is avail-
able. In addition, only one of the study health institu-
tions has recently started the service for patients with
MDR-TB
▸ Patients on re-treatment, who have experiences of
previous failures and usually have MDR-TB. When we
try to see “What happens to the depression at base-
line when the TB is treated?” we are assuming that
the treatment for the TB works well. Therefore, this
group of patients will be excluded.
SAMPLE SIZE DETERMINATION
The sample size was calculated using STATA V.12.0.33 We
used the following parameters: 80% power, 95% confi-
dence (one sided). We estimated that the prevalence of
treatment default among patients with TB without
depression will be 2.5% (estimate for the SNNPR).32 We
used an estimate of a 5% increase in prevalence of
defaulting among comorbid patients (ie, a prevalence of
treatment default of 7.5% among people with TB and
comorbid depression). Given the risk of death and drug
resistance among treatment defaulters, a 5% difference
could be a clinically important effect.34 We estimate that
we will find a ratio of 2:1 of non-exposed (not
depressed) to exposed (depressed) participants. With
these assumptions, the required sample size is 639. After
adding 10% contingency to account for potential loss to
follow-up, the total sample size is 703, of which 234 will
be exposed (with depression) and 469 will be non-
exposed (without baseline depression). In a recent
cross-sectional study carried out in outpatient healthcare
settings in Northern Ethiopia, the ratio of non-exposed
( patients with TB without depression) to exposed
( patients with TB with depression) was approximately
2:1. In the same study, 31% of the patients with TB had
depression, as measured using both SRQ-20 (cut-off 7 or
above) and Hospital Anxiety and Depression Scale
(cut-off eight or above) (Ambaw F. Convergence validity
of the Amharic-HADS among tuberculosis patients in
Bahir Dar and West Gojam Zone, Ethiopia(un published
manuscript). 2013).
3
Open Access

BMJ Open: first published as 10.1136/bmjopen-2015-007653 on 8 July 2015. Downloaded from http://bmjopen.bmj.com/ on May 4, 2022 by guest. Protected by copyright.
Figure 1 Conceptual framework of the study (BMI. body mass index; QOL, Quality of life; Rx, treatment; TB, tuberculosis).

RECRUITMENT AND DATA COLLECTION
The health professionals (usually BSc nurses and health
officers) running the TB clinic at each site will identify
the study participants who meet the inclusion criteria
consecutively until a total of 703 participants is reached,
and link them to experienced and trained diploma
holder nurse research assistants employed to collect data
full time. The health professionals in the TB clinics will
also have a role in helping the research assistants in
reviewing the patients’ charts for TB treatment out-
comes, comorbid illness and drug side effects. The
research assistants will provide information about the
study to the patient and seek informed written consent.
They will enrol patients who consent and arrange
appointment dates for the next interviews, and carry out
interviews at the health institutes. When patients do not

Figure 2 Map of the study area.

4 Ambaw F, et al. BMJ Open 2015;5:e007653. doi:10.1136/bmjopen-2015-007653

 Open Access

Table 1 Variables of the study and their plan of measurement

BMJ Open: first published as 10.1136/bmjopen-2015-007653 on 8 July 2015. Downloaded from http://bmjopen.bmj.com/ on May 4, 2022 by guest. Protected by copyright.
Measurement time
End of 2nd End of 6th Variable
Number Variables Baseline month month category
1 Depression √ √ √ Exposure
2 Treatment outcomes of TB (complete, cure, failure, √* √ Outcome
default, death, interruption)
3 Quality of life √ √ √ Outcome
4 Disability √ √ √ Outcome
5 Pathways to TB healthcare √ Outcome
6 Severity of signs and symptoms of TB √ √ √ Predictor
7 Sociodemographic variables √ Predictor
8 Substance use √ √ √ Predictor
9 Comorbid illness/negative life event √ √ √ Predictor
10 perceived social support √ √ √ Predictor
11 Perceptions of patients about TB √ Predictor
12 Medication side effects √ √ Predictor
13 Tuberculosis related stigma √ √ Predictor
√*=Only the presence/absence of treatment interruption, death and default will be measured.

consent to participate in the study, the research assistant
will request permission to record sex, age, level of educa-
tion and occupation of the patient. The list of variables
and timing of data collection is described in table 1.
MEASUREMENT
The primary exposure variable
Depression will be measured using PHQ-9.35 36 The
PHQ-9 has been validated and used in Ethiopia.37
Patients scoring 10 or more will be classified as having
high depressive symptoms (exposed). Symptoms will be
categorised using established cut-points 5, 10, 15 and 20
(mild, moderate, moderately severe and severe depres-
sion, respectively)35 36 to determine the different sever-
ities of exposure, and interval level scoring will be
computed to examine changes in intensity of depression
when the comorbid TB is treated.
Outcome variables
Primary outcome variable
Classical TB outcomes as defined by WHO28 and
Federal Ministry of Health of Ethiopia:9 treatment
defaulted, interrupted, failure, completed, cured or
death. Data on the patients’ status on the primary
outcome variable will be extracted from the TB register.
Secondary outcome variables
Quality of life (QOL): measured using a single item “How
would you rate your quality of life?”, with a continuous
numerical scale with values ranging from zero (worst

Figure 3 Schematic illustration for analysing the effect of tuberculosis on depression.

Ambaw F, et al. BMJ Open 2015;5:e007653. doi:10.1136/bmjopen-2015-007653 5

 Open Access
Figure 4 Schematic illustration for analysing the effect of depression on tuberculosis (QOL, Quality of life; TB, tuberculosis).
possible quality of life) to 10 (highest possible quality of
life) as recommended by de Boer et al38 and WHO.39
Disability: will be measured using the interviewer admi-
nistered version of the 12-item WHO Disability
Assessment Schedule V.2.0 (WHODAS V.2.0)40 with
scores ranging from 12 to 60, where higher scores will
represent worse levels of disability. WHODAS V.2.0 has
been used previously in Ethiopia and found to have con-
vergent validity.41 42
Pathways to healthcare: We are interested in finding out
which other care providers (ie, traditional healers, other
modern medicine providers) patients with TB have
visited in relation to symptoms related to the present
illness before attending the TB clinic. This will be mea-
sured at baseline using a modified version of the WHO
encounter form for mental disorders.43
Independent variables
Signs and symptoms of TB: Body mass index (weight in
kg/height in m2), fever, night sweats, cough, pain, per-
ceived weakness and anorexia will be measured using
single-item numerical indicator with scores ranging from
0 to 10, where 0 means absence of a symptom. The
items will be as follows: “How is your cough this week”,
“how is your fever this week”, “how are your night sweats
this week”, “how is your pain this week”, “how weak do
you feel this week”, and “how is your appetite this
week?”.
Sociodemographic variables: will include age, sex, marital
status, level of education, religion, ethnicity, average
household monthly income, perceived access to the
household income for the purpose of healthcare (no
access, only some access, adequate access), household
size, occupation, place of residence (urban vs rural), dis-
tance of the residence from health institute (estimates
in kilometres) and presence of dependent children.
Data on sociodemographic variables will be collected at
6 Ambaw F, et al. BMJ Open 2015;5:e007653. doi:10.1136/bmjopen-2015-007653
BMJ Open: first published as 10.1136/bmjopen-2015-007653 on 8 July 2015. Downloaded from http://bmjopen.bmj.com/ on May 4, 2022 by guest. Protected by copyright.
baseline using a structured questionnaire administered
by interviewers.
Substance use: The use of alcohol, tobacco and khat
among the study participants will be measured using the
Alcohol, Smoking and Substance Involvement Screening
Test (ASSIST) (V.3.1)44 at the three time-points. ASSIST
consists of eight items for each of alcohol and khat use
and seven items for tobacco products. It is a valid and
reliable scale in primary care settings in low-income and
middle-income countries.44–47
Comorbid illness and major life events: Data on the pres-
ence of comorbid illnesses such as HIV co-infection, car-
diovascular diseases, chronic obstructive pulmonary
diseases, diabetes mellitus and mental illnesses, includ-
ing previously diagnosed depression, will be collected by
asking the patient “Do you have another (other than
TB) diagnosed health problem(s)? Please specify”, at all
three time-points. If reported, the type of comorbid
illness, time of diagnosis and treatment being taken (if
any) will be specified by reviewing the participants’
medical records. Similarly, major life events of clients
over the last 2 months will be captured by asking the
question “Did you have an event (events) that you think
has negatively affected your life over the last 2 months
(eg, death of a family member, divorce, loss of
property?”.
Perceived social support: the individual’s evaluation of
whether and to what extent relationships are helpful48
and the expectation of whether support will be provided
within one’s social network.49 These two domains will be
measured using the three-item Oslo Scale of Perceived
Social Support,50 with scores ranging from 3 to 14, with
higher scores indicating better perceived social support.
Perceptions of patients about TB: perceived causes, sever-
ity, benefits of anti-TB treatment and barriers to anti-TB
treatment will be measured using open-ended semistruc-
tured interviews that will be coded thematically,

categorised and quantified so that responses may be
included in the quantitative analysis as categorical vari-
ables. The guiding questions will be ‘What do you think
causes TB?’ ‘How severe do you think TB can be?’ ‘How
helpful do you think the treatments of TB would be?’
‘What do you think your barriers may be to completing
your treatment’. Measurement will be taken at the
second month to minimise the burden of items at base-
line, when patients may be most ill.
Medication side effects: Information about medication
side effects will be extracted from the patient’s chart.
TB-related stigma: will be measured using a 10-item TB
stigma scale adopted from Macq et al51 in Nicaragua.
This instrument has been piloted on 68 patients with
TB, who will be excluded from the main study, and was
found to be understandable (except for one of the
items, which has been modified after the pilot), accept-
able to the respondents and data collectors, and to have
an internal consistency coefficient (α value) of 0.78.
The translation of the instrument into the local lan-
guage was made using the mixed methods translation
and consensus generation approach. This is a three-step
process. In step 1, panelists independently translate all
items and then independently rate every translation so
that the translations are categorised into inappropriate,
modifiable and appropriate. In step 2, modifiable trans-
lations are modified, rated independently and cate-
gorised into inappropriate or appropriate translation. In
step-three, a consensus discussion is made on items cate-
gorised as appropriately translated in step-one and
step-two, and ranked to get the best translations.52
TB-related stigma will be measured at the end of the
second month and at the end of the sixth month, to
provide time for social interaction after diagnosis and to
distinguish between stigmatising attitudes or behaviours
from appropriate precautions to prevent TB transmis-
sion. A patient with smear positive pulmonary TB
becomes non-infectious usually within 2–3 weeks except
in the case of drug resistance.9
Treatment for depression: For patients with depression
diagnosed by the health system, the appropriateness of
their treatments (for the depression) will be coded as
appropriate if they are in accordance with mhGAP
interventions guidelines and inappropriate if they do
not follow the recommended treatment approaches.
According to this guideline, an adult endorsing at least
two of the three core depression symptoms (depressed
mood, loss of interest in activities and decreased
energy), three other features of depression (reduced
concentration and attention, reduced self-esteem and
self-confidence, ideas of guilt and unworthiness, bleak
and pessimistic view of the future, ideas or acts of self-
harm or suicide, disturbed sleep, diminished appetite)
and having difficulties carrying out usual work, school,
domestic, or social activities over the same last 2 weeks
period in the absence of bereavement or other major
loss in prior 2 months, is identified as likely to have
moderate–severe depression.29
Ambaw F, et al. BMJ Open 2015;5:e007653. doi:10.1136/bmjopen-2015-007653
Open Access
The recommended treatment is: providing psychoedu-
BMJ Open: first published as 10.1136/bmjopen-2015-007653 on 8 July 2015. Downloaded from http://bmjopen.bmj.com/ on May 4, 2022 by guest. Protected by copyright.
cation, addressing existing psychosocial stressors, reacti-
vating social networks, a structured physical activity
programme, offering regular follow-up and considering
antidepressants (fluoxetine and amitriptyline, as well as
other tricyclic antidepressants).29 The nurses and health
officers of the health facilities, who are the main care
providers, have been trained by the health system and
the drugs are given to patients free of charge.
Data management
Data will be checked for consistency and completeness
by supervisors, double-checked by the principal investi-
gator and double entered to EpiData 3.1 by experienced
data entry clerks. Then, it will be exported to SPSS V.20
for cleaning and analysis. Hard copies of the data will be
stored in a locked cabinet and consent forms will be
separated from the data.
Data analysis
Data will be analysed using SPSS V.20. Descriptive statis-
tics will be computed to describe the sociodemographic
characteristics of participants and to summarise the dis-
tribution of each of the dependent (outcome) and inde-
pendent variables. We will check to see whether refusal
or losses to follow-up is associated with sociodemo-
graphic characteristics by testing the statistical signifi-
cance of differences between refusals and consenters,
and those lost to follow-up versus those retained at
6 months in terms of age, sex and level of education.
The prevalence of probable depression among
patients with TB at baseline will be determined by com-
puting the proportion of patients scoring 10 and above
on the PHQ-9 scale. The same analysis will be carried
out at endline and the magnitude of the difference
between the two measurements (baseline and endline)
will be calculated and the statistical significance tested
using McNemar test for repeated measures.
Determinants of depression at baseline will be examined
using logistic regression with depression as a dependent
variable, and sociodemographic characteristics, symp-
toms of TB, perceived social support and substance use
included as predictors. The proportion of patients with
TB that scored below 10 on PHQ-9 at baseline but who
had scores of 10 and above at the end of the second
month will be computed to determine incidence (risk)
of depression at 2 months. This will be repeated at
6 months.
The change in depressive symptoms (PHQ-9 scores)
over the three data collection points will be analysed
using multilevel mixed effect model by setting measure-
ment occasions as level 1 and individual patients with
TB as level 2 of the analysis. Level of disability, substance
use, presence of additional chronic illnesses, level of per-
ceived social support, perceptions of patients about TB
and appropriateness of depression treatment will be pre-
dictors at level-1; sociodemographic characteristics of
the individual will be predictors at level-2. We have
7
Open Access
selected this analysis approach as it does not require
complete data over measurement occasions, equal inter-
val of measurement for each case, or sphericity assump-
tions.53–56
The effect of depression on the treatment pathways of
patients with TB will be analysed using binary logistic
regression. The pathways followed by participants to
access care will be entered as separate-dependent vari-
ables; after coding each of the different pathways as zero
(if the specific pathway is not followed) or one (if the
specific pathway is followed), depression will be included
as an independent variable. We will adjust for sociode-
mographic characteristics.
To determine the effect of depression on the classical
outcomes of anti-TB treatment, logistic regression will
be carried out at the end of sixth months, with the treat-
ment outcomes (interrupted, defaulted, completed) as
dependent variables and depression as an independent
variable. We will adjust for substance use, stigma,
comorbid illnesses, sociodemographic factors, medica-
tion side effects, perceived social support and percep-
tions of the patient about TB. Although multinomial
logistic regression could handle the mutually exclusive
treatment outcomes as values of one dependent variable
with more than two categories, we have elected to use
logistic regression in order to simplify the interpretation
of our results.
The change in quality of life and disability levels over
time will be modelled by setting measurement occasions
as level-1 and individual patients as level-2. In this ana-
lysis, PHQ-9 scores, substance use, presence of additional
chronic illnesses, level of perceived social support,
appropriateness of depression treatment, level of TB
stigma, medication side effects and perceptions of
patients about TB will be modelled as level-1 predictors
and sociodemographic characteristics of the individual
will be predictors at level-2.
All the necessary assumptions will be tested and the
findings will be reported before inferential statistics are
calculated. Statistical tests will be considered significant
when p values are less than 0.05, and 95% CIs will be
presented throughout.
Ethical considerations
The proposal has been ethically cleared by the
Institutional Review Board (IRB) of College of Health
Sciences, Addis Ababa University, on 23 July 2014,
number 027/14/Psy. Informed written consent will be
obtained from every participant. Fingerprint impressions
will be taken from consenting illiterate participants.
Patients identified as having PHQ-9 scores 10 or above
or endorsing the suicide item of PHQ-9, will be referred
to nurses trained in mental health interventions by
another project working in collaboration with the
Ministry of Health of Ethiopia. We will give each partici-
pant 30.00 Birr ($1.50) to compensate them for the
time they spend with us. This amount of money is
8 Ambaw F, et al. BMJ Open 2015;5:e007653. doi:10.1136/bmjopen-2015-007653
enough to cover breakfast and a soft drink. We will not
BMJ Open: first published as 10.1136/bmjopen-2015-007653 on 8 July 2015. Downloaded from http://bmjopen.bmj.com/ on May 4, 2022 by guest. Protected by copyright.
give any other incentives.
Dissemination plan
Findings will be disseminated through publications in
peer-reviewed journals and conference presentations.
Summary reports will be submitted to the health institu-
tions and policymakers concerned. Community meet-
ings will be held to disseminate findings to the local
community, including study participants.
Limitations of the study
We have used PHQ-9 (a screening tool) to measure
depression, which may lead to error of categorising
patients into exposed and not exposed at baseline.
However, this will not compromise the work of capturing
the change in depression scores over time.
Currently, health institutes are actively strengthening
tracing mechanisms in order to decrease treatment
default and treatment interruption. This work may
mean that those who are otherwise more likely to
default or interrupt treatment may complete their treat-
ment, which may, in turn, weaken any association
between the exposure and these outcome variables.
There may also conceivably be a ‘treatment effect’ of
study participation in terms of improving depressive
symptoms and encouraging people to adhere to treat-
ment. In this low-income setting, people may have
undiagnosed comorbid illnesses and our method of cap-
turing comorbid illnesses may not be strong. Quality of
life will be measured using a single item and we will
therefore not obtain detailed information on the various
dimensions that make up this construct. In addition, our
conclusions cannot be applied to patients with MDR-TB
and patients on re-treatment for TB.
Strengths of the study
The study will provide much needed evidence about the
impact of comorbid depression on the course and
outcome of TB. The longitudinal study design will allow
us to estimate the incidence of depression among
people engaged with TB treatment. The observation of
depression TB comorbid patients from treatment initi-
ation to completion date will allow us to investigate
whether TB treatment alone may be sufficient to reduce
depressive symptoms. The study will enable us to investi-
gate the impact of depression in determining the course
and outcome of TB, independent of the effects of other
factors such as sociodemographic variables, stigma, per-
ceived social support, substance use and comorbid ill-
nesses such as HIV. As far as we are aware, this study will
be the first in Ethiopia, or in any African setting, to
examine the potentially important role of depression in
determining pathways to TB care.
Expected benefits of the findings
Our findings will contribute to a sparse evidence base
on mental health, TB and other chronic diseases in low-

income and middle-income countries. We hope that
other researchers will be encouraged to investigate
important questions in this area. We anticipate that our
study findings will contribute to raising awareness
among Ethiopian clinicians and service providers about
the potential impact of comorbid depression on the
course and management of chronic disease, as well as
informing future plans and policies related to the deliv-
ery of mental healthcare in primary care and other
healthcare settings in Ethiopia.
Author affiliations
1
School of Public Health, Bahir Dar University, Bahir Dar, Ethiopia
2
Department of Psychiatry, School of Medicine, College of Health Sciences,
Addis Ababa University, Addis Ababa, Ethiopia
3 disorders in TB/HIV co-infected patients in Ethiopia. BMC Infect Dis
Centre for Global Mental Health, Institute of Psychiatry, King’s College,
London, UK
Contributors FA, RM, CH and AA contributed equally to the design of the
study. FA drafted the manuscript and all the authors revised and approved the
manuscript.
Funding The research is funded by the European Union, Seventh Framework
Programme (FP7/2007–2013) under the Emerald grant agreement number
305968.
Competing interests None declared.
Ethics approval Institutional review board (IRB) of College of Health
Sciences, Addis Ababa University.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement The data will be available from the authors, with
permission from the Department of Psychiatry, Addis Ababa University.
Open Access This is an Open Access article distributed in accordance with
the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license,
which permits others to distribute, remix, adapt, build upon this work non-
commercially, and license their derivative works on different terms, provided
the original work is properly cited and the use is non-commercial. See: http://
creativecommons.org/licenses/by-nc/4.0/
REFERENCES
1. Katon WJ. Epidemiology and treatment of depression in patients
with chronic medical illness. Dialogues Clin Neurosci 2011;13:7–23.
2. Katon W, Lin EH, Kroenke K. The association of depression and
anxiety with medical symptom burden in patients with chronic
medical illness. Gen Hosp Psychiat 2007;29:147–55.
3. Ustun TB, Ayuso-Mateos JL, Chatterji S, et al. Global burden of
depressive disorders in the year 2000. Brit J Psychiatr
2004;184:386–92.
4. WHO. The world health report: mental health: new understanding,
new hope. 2001.
5. WHO. The effectiveness of mental health services in primary care:
the view from the developing world. 2001;WHO/MSD/MPS/01.1.
6. Murray CJ, Vos T, Lozano R, et al. Disability-adjusted life years
(DALYs) for 291 diseases and injuries in 21 regions, 1990–2010:
a systematic analysis for the Global Burden of Disease Study 2010.
Lancet 2012;380:2197–223.
7. Hailemariam S, Tessema F, Asefa M, et al. The prevalence of
depression and associated factros in Ethiopia: findings from the
National Health Survey. Int J Ment Health 2012;6:23.
8. Abdulahi H. Burden of disease analysis in rural Ethiopia. Ethiop Med
J 2001;39:271–81.
9. Federal Ministry of Health of Ethiopia. Guidelines for clinical and
programmatic managment of TB, Leprosy and TB/HIV in Ethiopia.
5th edn. Addis Ababa: Falcon Printing, 2012.
10. WHO. Global tuberculosis control report. 2011.
11. Issa BA, Yussuf AD, Kuranga SI. Depression comorbidity among
patients with tuberculosis in a university teaching hospital outpatient
clinic in Nigeria. Ment Health Fam Med 2009;6:133–8.
Ambaw F, et al. BMJ Open 2015;5:e007653. doi:10.1136/bmjopen-2015-007653
Open Access
12. Panchal SL. Correlation with duration and depression in TB patients
in rural Jaipur district (NIMS Hospital). Int J Pharm Bio Sci
BMJ Open: first published as 10.1136/bmjopen-2015-007653 on 8 July 2015. Downloaded from http://bmjopen.bmj.com/ on May 4, 2022 by guest. Protected by copyright.
2011;2:263–7.
13. Sulehri MA, Dogar IA, Sohail H, et al. Prevalence of depression
14.
among tubercolosis patients. APMC 2010;4:133–7.
Aydin IO, Ulusahin A. Depression, anxiety comorbidity, and disability
in tuberculosis and chronic obstructive pulmonary disease patients:
applicability of GHQ-12. Gen Hosp Psychiatry 2001;23:77–83.
15. Iga OM, Lasebikan VO. Prevalence of depression in tuberculosis
patients in comparison with non-tubercolosis family contacts visiting
the DOTS clinic in a Nigerian tertiary care hospital and its correlation
with disease pattern. Ment Health Fam Med 2011;8:235–41.
16. Trenton AJ, Currier GW. Treatment of comorbid tuberculosis and
depression. Prim Care Companion J Clin Psychiatry 2001;3:
236–43.
17. Moussas G, Tselebis A, Karkanias A, et al. A comparative study of
anxiety and depression in patients with bronchial asthma, chronic
obstructive pulmonary disease and tuberculosis in a general hospital
of chest diseases. Ann Gen Psychiatry 2008;7:7.
18. Deribew A, Tesfaye M, Hailmichael Y, et al. Common mental
2010;10:201.
19. Simon GE, Von Korff M, Lin E. Clinical and functional outcomes of
depression treatment in patients with and without chronic medical
illness. Psychol Med 2005;35:271–9.
20. Sweetland A, Oquendo M, Wickramaratne P, et al. Depression:
a silent driver of the global tuberculosis epidemic. World Psychiatry
2014;13:325–6.
21. Mikkelsen RL, Middelboe T, Pisinger C, et al. Anxiety and
depression in patients with chronic onstructive pulmonary disease
(COPD): A review. Nord J Psychiat 2004;58:65–70.
22. Vega P, Sweetland A, Acha J, et al. Psychiatric issues in the
management of patients with multidrug-resistant tuberculosis. Int J
Tuberc Lung Dis 2004;8:749–59.
23. Reiche EM, Nunes SO, Morimoto HK. Stress, depression, the
immune system, and cancer. Lancet Oncol 2004;5:617–25.
24. Kiecolt-Glaser JK, Glaser R. Depression and immune function:
central pathways to morbidity and mortality. J Psychosom Res
2002;53:873–6.
25. Rayner L, Price A, Evance A, et al. Antidepressants for depression
in physically ill people. Cochrane Database Syst Rev 2010;(3):
CD007503.
26. Rodin G, Voshart K. Depression in the medically ill: an overview.
Am J Psychiat 1986;143:696–705.
27. Doherty AM, Kelly J, McDonald C, et al. A review of the interplay
between tuberculosis and mental health. Gen Hosp Psychiat
2013;35:398–406.
28. WHO. Treatment of tuberculosis: guidelines. 4th edn. WHO/HTM/
TB/2009.420. 2010.
29. WHO. Mental health gap action programme intervention guide for
mental, neurological and substance use disorders in non-specialized
health settings. 2010. http://whqlibdoc.who.int/publications/2010/
9789241548069_eng.pdf (accessed Dec 2012).
30. Fedeal Ministry of Health of Ethiopia. Ethiopian mental health
strategy. 2012. http://www.centreforglobalmentalhealth.org/sites/
www.centreforglobalmentalhealth.org/files/uploads/documents/
ETHIOP~2.pdf (accessed Jan 2013).
31. WHO. Global tuberculosis report. 2012.
32. Federal Democratic Republic of Ethiopia Ministry of Health. Health
and Health Related Indicators. 2011. http://www.dktethiopia.org/
sites/default/files/PublicationFiles/Health%20and%20Health%
20Related%20Indicators%202003%20E.C.pdf (accessed Dec 2013).
33. StataCorp LP. USA. 2011. http://www.stata.com/stata11/
34. Ferguson CJ. An effect size primer: a guide to clinicians and
researchers. Prof Psychol Res Pr 2009;40:532–8.
35. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief
depression severity measure. J Gen Intern Med 2001;16:606–13.
36. Kroenke K, Spitzer RL, Williams JB, et al. The patient health
questionnaire somatic, anxiety, and depressive symptom scales:
a systematic review. Gen Hosp Psychiat 2010;32:345–59.
37. Gelaye B, Williams MA, Lemma S, et al. Validity of the patient health
questionnaire-9 for depression screening and diagnosis in East
Africa. Psychiat Res 2013;210:653–61.
38. de Boer AG, van Lanschot JJ, Stalmeier PF, et al. Is a
single-item visual analogue scale as valid, reliable and responsive
as multi-item scales in measuring quality of life? Qual Life Res
2004;13:311–20.
39. WHO Regional Office for Europe (EUROHIS). Developing common
instruments for health surveys. 2003.
40. WHO. Measuring Health and Disability: Manual for WHO Disability
Assessment Schedule, WHODAS 2.0. 2010. http://whqlibdoc.who.
9
 Open Access
int/publications/2010/9789241547598_eng.pdf (accessed 23 Oct
2013).
41. Mogga S, Prince M, Alem A, et al. Outcome of major depression in
Ethiopia: population-based study. Br J Psychiatry 2006;189:241–6.
42. Senturk V, Hanlon C, Medhin G, et al. Impact of perinatal somatic
and common mental disorder symptoms on functioning in Ethiopian
women: the P-MaMiE population-based cohort study. J Affect Disord
2012;136:340–9.
43. WHO. Pathways of patients with mental disorders. 1987. http://
whqlibdoc.who.int/hq/1987/MNH_NAT_87.1.pdf (accessed 24
Nov2013).
44. WHO. The Alcohol, Smoking and Substance Involvement Screening
Test (ASSIST): Manual for use in primary care. 2010.
45. Humeniuk R, Ali R, Babor TF, et al. Validation of the Alcohol,
Smoking and Substance Involvement Screening Test (ASSIST).
Addiction 2008;103:1039–47.
46. Newcombe DA, Humeniuk RE, Ali R. Validation of the World Health
Organization Alcohol, Smoking and Substance Involvement
Screening Test (ASSIST): report of results from the Australian site.
Drug Alcohol Rev 2005;24:217–26.
47. WHO ASSIST Working Group. The Alcohol, Smoking and
Substance Involvement Screening Test (ASSIST): development,
reliability and feasibility. Addiction 2002;97:1183–94.
48. Schaefer C, Coyne JC, Lazarus RS. The health-related functions of
social support. J Behav Med 1981;4:381–406.
10
49. Seidlecki KL, Salthouse TA, Oishi S, et al. The relationship between
social support and subjective well-being across age. Soc Indic Res
BMJ Open: first published as 10.1136/bmjopen-2015-007653 on 8 July 2015. Downloaded from http://bmjopen.bmj.com/ on May 4, 2022 by guest. Protected by copyright.
2014;117:561–76.
50. Meltzer H. Development of a common instrument for mental health.
In: Nosikov A, Gudex C, eds. EUROHIS: developing common
instruments for health surveys. Amsterdam: IOS Press, 2003.
51. Macq J, Solis A, Martinez G, et al. Tackling tuberculosis patients’
internalized social stigma through patient centred care: an
intervention study in rural Nicaragua. BMC Public Health
2008;8:154. doi:10.1186/1471-2458-8-154
52. Sumathipala A, Murray J. New approach to translating instruments
for cross-cultural research: a combined qualitative and quantitative
approach for translation and consensus generation. Int J Meth Psych
Res 2000;9:87–95.
53. Blackwell E, de Leon CF, Miller GE. Applying mixed regression
models to the analysis of repeated-measures data in psychosomatic
medicine. Psychosom Med 2006;68:870–8.
54. Burton P, Gurrin L, Sly P. Extending the simple linear regression
model to account for correlated responses: an introduction to
generalized estimating equations and multi-level mixed modelling.
Stat Med 1998;17:1261–91.
55. Heck R, Thomas S, Tabata L. Multilevel and longitudinal modeling
with IBM SPSS. 2nd edn. USA: Taylor & Francis, 2014.
56. Tabachnic BG, Fidell LS. Using multivariate statistics. 5th edn:
Pearson Educ. Inc., 2007.
Ambaw F, et al. BMJ Open 2015;5:e007653. doi:10.1136/bmjopen-2015-007653